NWD: Since this is an interview and not a literature review, we would like to start with getting to know you …
Dr. M: I graduated from Case Western Reserve University School of Dentistry, Cleveland, Ohio, in 1983. Not knowing what I wanted to do at that time, I worked as an associate in a large dental office for a couple of years as a general dentist. During that time I explored the various specialties. A friend of mine, who was the group practice’s periodontist, piqued my interest in this field. Periodontology was a good fit for both my clinical and science interests; it seemed to be one of the more science-driven disciplines in dentistry. I interviewed at a number of schools, including the University of Minnesota.
Among the first people I met here were Bruce Pihlstrom, who directed the graduate program, and Erwin Schafer, a periodontist and former dean of the School. I enrolled in the graduate program in 1985 and completed my Master’s thesis in 1988. I had planned to return to Michigan, where I grew up, and enter private practice. Because Michigan had a specialty exam that was only given once a year, I had to wait until the following year to take this exam. In the meantime, Bruce Pihlstrom cobbled together some research funding and offered me a position. By the time I took the state exam — which I passed — I was entrenched in several of Bruce’s research projects and decided to stay at Minnesota. I also continued to work on a study of genetic and environmental influences on periodontitis, which was the subject of my graduate thesis. Bruce had obtained a National Institutes of Health (NIH) grant to study rearedtogether and reared-apart twins. I first served as a co-investigator, then principal investigator, on this project. At the time, it was one of few ongoing twin studies in our field, although it wasn’t the first.
In 2002, Bruce left the University (and literally this desk!) to become director of the clinical research division at the National Institutes of Dental and Craniofacial Research (NIDCR). He had been the first and only Erwin Schafer Chair holder. I was offered the Schafer Chair, and also became Director of the Oral Health Clinical Research Center. Before he left, Bruce and others were pursuing funding for a study to determine if periodontal treatment improved rates of pre-term birth and other adverse pregnancy outcomes. I assumed the leadership role in that study, which was subsequently funded by the NIDCR. Since then I have been researching various oral health and general health associations, including whether treatment of periodontal disease can alter the progression or the risk for pre-term birth and type 2 diabetes.
NWD: It sounds like the people you found here were a major influence in your life.
Dr. M: Times were different when I started as a graduate student. We had a larger and more diverse faculty in the division than we do today. This afforded us the luxury to focus concurrently on research, teaching, and service. It was a vibrant atmosphere. I wanted to be part of that group and benefit from their knowledge, experiences, and willingness to mentor junior faculty
That said, today Periodontology remains a strong division with very solid education programs, especially at the post-doctoral level. We strive to keep our pre-and post-doctoral curricula and clinical instruction evidence-based. While much has changed, I still believe that periodontology remains strongly rooted in science.
NWD: Tell us about the Schafer Chair — any perks?
Dr. M: I think it is still the only endowed chair in dentistry, and over the years the endowment has grown nicely. It was created in 1986 or 7; Carl Bandt chaired the fundraising campaign. The campaign raised $1 million, an amount that was matched by the University. The relatively short time it took to raise the money was a testimonial to Erwin’s standing in the community. The endowment’s “principal” remains with the University Foundation. The investment return on this money is used to grow the endowment and to support activities of the Chair. We use the Schafer Chair to provide seed money for graduate student research and to fund pilot projects that we hope will lead to external funding. The Chair also supports that fraction of my salary not recovered by research grants.
I have to say that having that title (Schafer Chair) under my name is a constant reminder of Erwin’s legacy, enthusiasm, and commitment to education. I try to be the strongest steward I can of that resource.
NWD: There is a great deal of new information purportedly linking periodontal disease to serious consequences in our general health; thus the not always facetious feeling that “perio causes everything” that was the inspiration for this interview.
Dr. M: An important issue in this field of research is to determine the specific nature of the associations. Periodontal disease is strongly associated with higher age and lower education and income and with smoking or tobacco use. Thus, diseases that are more prevalent in these groups (the elderly or smokers) will be associated with periodontal disease because they have shared risk factors and not necessarily because periodontitis somehow affects risk for the medical disorder. There needs to be some biological reason why the two conditions are linked. For example, do the organisms that cause periodontal disease play a role in the pathogenesis of these other conditions (diabetes, cardiovascular disease, pre-term birth), or is their presence simply a good indicator of who may be at risk for these other conditions? There are a number of criteria in epidemiology that must be met before something can be labeled as a “risk factor” for something else. For example, there should be evidence of dose-dependency (that the risk for the medical disorder is highest in those with the most severe periodontitis). We must also establish the temporal sequence of events. In other words, if periodontal disease is thought to “cause” another condition, it must precede the onset of the medical condition. Also, periodontal treatment, by reducing the patient’s exposure to the disease, should reduce their risk for the related condition.
NWD: Was the low-birth-weight investigation the first?
Dr. M: I think the cardiovascular disease-periodontal disease link was the first, but even going back to the early 20th century, oral infections were thought to cause diseases in distant organs. This was the so-called “focal infection” theory, which was debunked, and for decades there was no discussion about periodontal disease causing non-oral conditions. In the late 80s/early 90s, a number of studies established a link between periodontal and cardiovascular disease; the first solid evidence linking periodontal disease to pre-term birth risk was reported by Steve Offenbacher and colleagues in 1996.
NWD: Here’s the question with the long answer — let’s address the research.
Dr. M: First, as health care professionals, I think it is incumbent on us to research what we “hear” in the lay press or in conversations with others by delving into the scientific literature. If somebody tells us, for example, that periodontal disease causes cardiovascular disease, our scientific curiosity and sense of commitment to our patients and the public should compel us to learn for ourselves if the evidence is as strong as is being purported. If the evidence is solid, we should ask, “What are the clinical implications of this information?” My wife reads a number of women’s journals and most months she reads to me a blurb about oral health or about some new oral health care device. Most of the time, the piece lacks detail or study references and makes sweeping statements about the finding’s applications. Because the public is inundated with unfiltered health care information, we as professionals must be diligent about educating our patients and making treatment decisions or recommendations that are soundly based in science.
As for oral health/general health connections, many medical conditions have “passed” the initial test in that they show associations with periodontal disease on a crosssectional basis. Some associations, like the one between periodontitis and cardiovascular disease, have been confirmed in prospective studies, but many have not. (In a prospective study, groups of individuals with or without periodontal disease at the start of the study are followed over time to determine if the rate of development of the other condition differs between the groups.) Studies on everything from gastric and head and neck cancers to erectile dysfunction to Parkinson’s disease are being conducted, but many are in the very early stages of investigation and, as I have alluded to earlier, there are a number of reasons why two conditions may be associated statistically, but not in a cause-and effect m
Low-grade inflammation has been associated with an increasing number of disorders, including some cancers and certainly cardiovascular disease. Because periodontal disease is essentially a low-grade inflammatory disorder, many researchers have postulated that molecules that regulate inflammation and that are generated within the periodontal tissues travel through the vasculature and affect the risk for or progression of other medical disorders. It is almost too easy in this regard to conclude that most oral health-general health associations are biologically plausible because of a shared “inflammatory component”. The specific role of periodontal pathogens or of periodontal-derived inflammation on other organ systems must be more clearly defined before a cause-and-effect relationship can be inferred. Some of these mechanistic studies must be done in animals, although how well a particular animal model mimics the human condition is always of concern. One generally untested theory is that periodontal disease may be linked to another inflammatory condition because both are manifestations of an “inflammatory phenotype”. That is, some individuals, through either genetic predisposition and/or environmental exposures, are more likely to develop strong inflammatory responses or are more prone to the effects of inflammation, regardless of its cause.
Another issue to consider when evaluating the evidence is the strength of the association. It is important not to confuse statistical significance with the actual strength or magnitude of the association. Headlines in the press or even in scientific journals may be based on the former and not the latter. With regard to the periodontal disease/cardiovascular disease link, associations have been reported in both crosssectional and prospective studies, but the associations are not particularly strong. From a practical standpoint, the strength of the association can affect how we apply that information in a clinical setting. If the association is weak — even if it is highly statistically significant — then periodontal disease may explain only a small portion of the risk for the “related” condition. In other words, many periodontally healthy patients may develop the “related” condition while many patients with periodontal disease may remain unaffected. Also, all of the conditions associated with periodontal disease to date — cardiovascular disease, diabetes, poor pregnancy outcomes, for example — are multifactorial, meaning many factors contribute to the disease risk and progression.
Another issue that should come up in this discussion is that of “confounding”. A good example is cigarette smoking, which is associated with both periodontal disease and cardiovascular disease. Thus, periodontal disease may be associated with cardiovascular disease — both cross-sectionally and prospectively — because of the role of smoking in both conditions. It may be difficult, if not impossible, to disentangle the effects of confounders in the statistical treatment of data. One of the “cleanest” ways to eliminate confounding is to explore associations in groups of individuals who are not exposed to the confounder. Few studies have been large enough to study associations in subgroups of individuals such as non-smokers. Among the few that could, some have reported weaker associations between periodontal disease and cardiovascular disease among non- smokers, whereas others have reported similar associations in this subgroup.
If the science supports an association between periodontal disease and a medical disorder, it is reasonable for practitioners to ask, “What can I do to help reduce my patient’s risk for that disorder?” I think it is interesting in this regard to examine the evolution of the science in terms of the relationship between periodontal disease and poor pregnancy outcomes (pre-term birth/low-birth-weight babies). The association between a woman’s periodontal health and her risk for a poor pregnancy outcome is consistent and fairly modest. Overall, studies suggest that women with periodontitis are about two to three times more likely than periodontally healthy women to have a pre-term infant. The question is whether we as care providers can affect this risk. Several early studies suggested that periodontal treatment in pregnant women reduced the risk for pre-term birth.
These studies, however, were conducted at a single center and were relatively small in terms of the numbers of women with pre-term births. Since then, at least five large studies have provided us with a more definitive answer. They have consistently shown that scaling and root planing do not reduce the rate of pre-term birth or low birth weight in women with periodontitis. This is an example of why I believe it is important to wait until high-quality evidence is available before making patient care decisions or promising our patients more than we can deliver. Before these large studies were completed, many groups, including the March of Dimes, promoted periodontal treatment as a means of reducing pre-term birth rates in the U.S. There are several important and unanswered questions, such as whether or not delivering treatment earlier in pregnancy or even before conception can improve birth outcomes, but for a pregnant woman with periodontitis, we should not be promoting our treatment as a means to reduce her chance of having an unfavorable birth outcome.
As a dentist, I continue to put a premium on periodontal health, on disease prevention, and on providing treatment to improve oral health per se. Despite the myriad studies linking periodontal disease to other medical conditions, I do not know of any solid evidence to suggest that periodontal treatment can alter one’s risk for any medical condition. On the flip side of this, I also know of no studies that show periodontal treatment causes harm or raises one’s risk for pre-term birth, diabetes, cardiovascular disease, etc.
Currently, I am working with a large group to explore the relationship between periodontal disease and type 2 diabetes. I think the evidence is clear that diabetes, particularly long-standing or poorly-controlled disease, is a risk factor for periodontal disease. There is some evidence that periodontal disease may increase one’s risk for developing diabetes and that periodontal treatment may improve blood sugar control in patients with type 2 diabetes. Just as we discussed earlier, though, it is important to conduct large-scale studies to determine if this treatment effect is evident in diverse populations. Such studies are needed to help establish if a cause-and-effect relationship exists and whether we as therapists can alter the risk or progression of the disease by what we do in the mouth. We hope to complete enrollment of this multicenter trial in mid-2012 and have the results available in early 2013.
NWD: How much does the result of one study impact what happens next throughout the research community?
Dr. M: I am concerned that people, even health care professionals, are too willing to accept as gospel the results of an initial study. How confident am I that findings from a single small study pertain to my patient in the chair? Reproducibility and generalizability are always concerns in science. With regard to oral healthgeneral health associations, I do not want to alarm my patients based on something that they or I heard and that may not have been confirmed or corroborated. Unfortunately, dentistry tends to rely on smaller, shorter-term studies than does medicine, and we tend to take that leap of faith that all study findings are real and pertain directly to our patient populations.
Fortunately, the projects that I have been involved with in this area have been funded by the NIH and, as such, have been vetted through the scientific review process. The search for disease associations might be viewed as “hypothesis-generating exercises”. Clinical trials to further explore these associations are “hypothesis-testing exercises”. This is an important distinction because everything — the sample population, the treatment provided, the outcome of interest, even the data analysis plan — must be pre-specified in a clinical trial. Such an approach prevents us from collecting mounds of data, and then, as statisticians warn against, “torturing the data until it confesses”. Such an unstructured approach to data analysis is likely to lead to false positive (or un-reproducible) findings.
On a related note, historically there had been such concern about investigators or industry suppressing unfavorable clinical trials data or reporting only on outcomes that showed a treatment effect, that most top-tier medical (and some dental) journals now require investigators to report, in the public domain, their detailed study plans before the trial commences. There are several clinical trials registries throughout the world. One is clinicaltrials.gov, which is maintained by the NIH. Oral health care providers who are interested in learning if a new product or device or procedure is being tested can search these websites using common search terms. Recently, search under the term “dentistry” yielded 600 studies; more than 3,800 were accessible through the phrase “oral health”.
NWD: Once you define the design, do you get the time you need?
Dr. M: As I mentioned, I have been fortunate to have been funded for these projects through the NIH. So a portion of my salary (my time) is paid by the federal government to work on these projects. The NIH is interested in having its funded investigators succeed, and it understands that it takes a number of skilled individuals to complete a trial. In that regard, they try to make available the resources needed to meet the support needed by your study team. During the grant writing process, one typically has to “find” the time to write (and rewrite) the grant. But if the project is funded, investigators are able to dedicate a portion of their time to the study. The NIH actually insists on this!
As for the specific study design, our work of late has been focused on treating patients and monitoring them for other health outcomes (pre-term birth in pregnant patients, blood sugar control in patients with type 2 diabetes). As I mentioned earlier, clinical trials are needed to more thoroughly explore the nature of oral health-general health associations, but they also serve as the basis for what we do in practice. It is interesting to see how often what people think they know is refuted by randomized, controlled, and blinded clinical trials. Great examples can be found in the Women’s Health Study (regarding estrogen use) or the ACCORD Study (regarding the desired level of glycemic control in patients with type 2 diabetes) where “conventional wisdom” and common clinical practices were, in some sense, turned upside down.
As the School of Dentistry and the field places a greater emphasis on evidence-based health care, we need to teach our students more about the quality and amount of evidence they should demand in order to guide their patient care decisions. Although there is an “art and science” of dentistry, the science should come first, and I believe dentists really want solid direction from the science. With regard to oral health-general health links, what do we tell the practitioner? I would tell him or her to place great value on oral health regardless of whether or not it affects the patient’s general medical health. We still teach our students to provide periodontal care and preventive services to improve (or maintain) the patient’s oral function, comfort, and esthetics. If in doing that we can positively affect other health outcomes, wonderful; that is icing on the cake, but it is not the main reason we are providing that care nor should it be used to “scare” patients into accepting treatment.
NWD: So how about that “perio causes everything”?
Dr. M: [laughs] I had a lecture slide that listed conditions that had been associated with periodontal disease. I kept updating it until the list included about 30 conditions. The font got so small that students couldn’t read it regardless. Even measures of balance and physical fitness have been associated with periodontal health. Many wide-ranging conditions and disorders have shown associations with periodontal disease. Inflammation seems to be common link for many. As I mentioned, it is possible that some individuals are “hyperresponders” and thus prone to multiple inflammatory conditions, including periodontitis. We simply don’t know yet. I think the strongest and most consistent evidence in support of a causal link is between periodontal disease and cardiovascular disease. As I mentioned, though, the magnitude of the association, though consistent, isn’t particularly strong. Importantly, no study has found that treatment of periodontal disease reduces the risk for any cardiovascular disease. To date, I don’t believe that periodontal disease has been shown to be a risk factor (i.e., have a direct cause-and-effect relationship) with any medical condition or disorder. That is not to say that oral bacteria or uncontrolled oral infections cannot have adverse health effects. A clear example comes from the recent obstetrics literature in which an oral organism was implicated as the cause of a term stillbirth. The evidence in this case strongly suggested that a strain of Fusobacterium nucleatum, present in the subgingival plaque, invaded the fetus and lead to its demise. Interestingly, this patient had gingivitis but not periodontitis. While such reports reinforce the notion that the mouth is indeed connected to the rest of the body, they do not support the concept that periodontal disease is a causative factor, in this case, for stillbirths in the general population or that periodontal treatment can reduce these risks.
NWD: What do you think about the future direction of oral health research?
Dr. M: I value the importance of basic science research and certainly clinical research. That said, as I think about the problems and challenges that our patients now face, I increasingly wonder about the impact that future discoveries will have on our patients. Research and discovery is important, critically important, but we need to focus equal or greater resources on finding better and more efficient ways to translate current knowledge and technologies into preventive and patient care services. While it is important to discover the next inflammatory biomarker or oral health-general health link, the public will likely be best served by finding ways to help patients better prevent oral disease and to improve access to utilization of oral care services. We need to apply the same rigorous scientific principles we have to the basic and clinical sciences for so long to find solutions to these problems as well.